US20250021309
2025-01-16
Physics
G06F8/20
The patent application describes a system and method for generating web or mobile applications through a front-end graphics editor. This editor allows users to design user interfaces (UI) visually, and employs artificial intelligence (AI) to generate corresponding human-readable front-end code. The system further utilizes a large language model (LLM) transformer to convert this code into a machine-readable format, facilitating the construction of the application.
Recent advancements in automated tools have significantly simplified web and mobile application development, particularly in UI design. Platforms like Figma enable rapid prototyping by using pre-defined components, yet manual code generation for UI elements remains labor-intensive. AI tools such as GitHub Copilot have emerged, offering extensive code suggestions by leveraging vast datasets of public code. However, integrating front-end and back-end code remains a challenge.
The system supports collaborative environments where multiple designers can work simultaneously on a UI design. It allows for importing design files from platforms like Figma and generates human-readable code for each UI element, ready for back-end integration. The UI is abstracted into a view-controller object that is environment-independent, facilitating seamless integration with back-end systems.
The method includes defining variant components with multiple states and interactions, enabling dynamic UI behavior at runtime. These components are represented on tool panels for easy manipulation. The generated code supports conversion into language-specific objects suitable for various runtime environments, ensuring compatibility across platforms.